Northern Territory youth justice models. Northern Territory youth justice models Fixing a broken system. 24 October 2017

Transcription

1 \ Northern Territory youth justice models Fixing a broken system 24 October

2 Contents Glossary Key terms Executive summary v vi vii 1 Background Overview of modelled intervention Key assumptions and caveats 20 2 Base case Base case model Base case trajectories Cost inputs Base case results 33 3 Intervention Diversion Bail accommodation Restorative detention 48 4 Cost benefit analysis Summary of intervention and results Combined costs and benefit analysis Summary of results Sensitivity analysis Conclusion 78 References 79 Appendix A: Detailed tables 83 A.1 Demographics and apprehensions 83 A.2 Supervised youths in the base case 86 Limitation of our work 90 General use restriction 90 i

3 Charts Chart i : Breakdown of total benefits following intervention xii Chart ii : Results of sensitivity analysis, BCR and net benefit, real dollars, $ millions xiii Chart 1.1 : Young people aged in detention on an average day, by legal status and jurisdiction, (rate) 15 Chart 2.1 : Youth population by Indigenous status in the NT, 2007 to Chart 2.2 : Youth population by gender in the NT, 2007 to Chart 2.3 : Youth population by age in the NT, 2007 to Chart 2.4 : Youths apprehended and rate of apprehensions per distinct youth in the NT, by year 27 Chart 2.5 : Historical and projected offence profile for young people in detention in the NT, total and relative to population 27 Chart 2.6 : Projected numbers of supervised youth in the NT 29 Chart 2.7 : Costs of detention and community-based supervision per youth per day in the NT (real dollars) 32 Chart 3.1 : Six month re-conviction rates under regular and restorative detention in Washington DC. 58 Chart 3.2 : Ratios of recidivism after restorative detention to recidivism after regular detention, by years since release (US) 60 Tables Table i : Modelling inputs and assumptions ix Table ii : Overall costs and benefits of diversion intervention, real discounted dollars (rate = 7%) x Table iii : Overall costs and benefits of bail accommodation, real discounted dollars (rate = 7%) x Table iv : Unique youths sentenced to detention after diversion and bail accommodation, no restorative detention xi Table v : Overall costs and benefits of restorative detention, real discounted dollars (rate = 7%) xi Table vi : Overall costs and benefits of intervention, real discounted dollars (rate = 7%) xii Table 1.2 : Data requested for modelling 20 Table 1.3 : Key assumptions and limitations used in the modelling 21 Table 2.1 : Supervised youths by year in the NT 30 Table 2.2 : Number of participants by diversion category, to Table 2.3 : Cost of crime, inputs by cost type 31 Table 2.4 : Average length of stay for supervision types in the NT (days) 33 Table 2.5 : Supervised youth and days of supervision in the NT, by year 34 Table 2.6 : Costs of supervision in the NT, current year, $ million, dollars 34 Table 3.1 : Cost per person, dollars 36 Table 3.2 : Juvenile diversion rates and number of apprehensions by crime type, non-indigenous, Table 3.3 : Overall costs of diversion, real dollars. 38 Table 3.4 : Recidivism by diversion program 40 Table 3.5 : Change in diversion participants and associated reduction in apprehensions and days of supervision in the NT 42 Table 3.6 : Change in adult incarcerations and costs of crime and incarcerations due to increased diversion expenditure, real dollars, undiscounted 43 Table 3.7 : Overall costs and benefits of diversion intervention, real discounted dollars (rate = 7%) 44 ii

4 Table 3.8 : Overall costs of bail accommodation 45 Table 3.9 : Participants and change in apprehensions following bail accommodation 46 Table 3.10 : Change in the number of supervised youth and funding and associated reduction in days of supervision due to bail accommodation 47 Table 3.11 : Change in adult incarcerations and costs of crime and incarcerations due to bail accommodation, real dollars, undiscounted 48 Table 3.12 : Overall costs and benefits of bail accommodation, real discounted dollars (rate = 7%) 48 Table 3.13 : Total capital costs for proposed restorative detention facilities 53 Table 3.14 : Number, occupation and salaries of professional staff in MST facilities, and proposed NT facilities 55 Table 3.15 : Number, occupation and salaries of professional staff in MST facilities, and equivalent NT facilities 56 Table 3.16 : Observed differences in recidivism between regular and restorative detention across US States (various years) 59 Table 3.17 : Proportion of offenders with prior incarceration, by jurisdiction, Table 3.18 : Number and proportion of youths returning to sentenced detention within 12 months, Australia, to Table 3.19 : Unique youths sentenced to detention without diversion and bail accommodation, no restorative detention 63 Table 3.20 : Unique youths sentenced to detention after diversion and bail accommodation, with restorative detention 64 Table 3.21 : Change in detention population and associated reduction in days of supervision 65 Table 3.22 : Change in adult incarcerations and costs of crime and incarcerations due to restorative detention, real dollars, undiscounted 65 Table 3.23 : Overall costs and benefits of restorative detention, real discounted dollars (rate = 7%) 66 Table 4.1 : Modelling inputs and assumptions 67 Table 4.2 : Overall costs and benefits of diversion intervention, real discounted dollars (rate = 7%) 68 Table 4.3 : Overall costs and benefits of bail accommodation, real discounted dollars (rate = 7%) 68 Table 4.4 : Overall costs and benefits of restorative detention, real discounted dollars (rate = 7%) 69 Table 4.5 : Overall costs of intervention, real dollars 69 Table 4.6 : Change in intervention participants and associated reduction in apprehensions and days of supervision 70 Table 4.7 : Change in adult incarcerations and costs of crime and incarcerations due to intervention, real dollars, undiscounted 71 Table 4.8 : Overall costs and benefits of intervention, real dollars 71 Table 4.9 : Net benefit of intervention with lower detention cost growth, real dollars 72 Table 4.10 : Net benefit of intervention with varying levels of diversion, real dollars 73 Table 4.11 : Net benefit of intervention with higher average cost of diversion programs, real dollars 73 Table 4.12 : Net benefit of intervention with higher reoffending rates for diversion programs, real dollars 74 Table 4.13 : Net benefit of intervention with higher average daily costs of bail accommodation, real dollars 74 Table 4.14 : Net benefit of intervention with higher average daily costs of restorative detention, real dollars 75 Table 4.15 : Net benefit of intervention with varying recidivism rates, real dollars 76 Table 4.16 : Net benefit of intervention with varying discount rates, real dollars 76 Table 4.17 : Sensitivity analysis results, real discounted dollars, $ millions 77 iii

5 Figures Figure 1.1 : Summary of modelled interventions in the justice system 20 Figure 2.1 : Average pathways of youths apprehended in the NT, to Figure 3.1 : Layout of the Northern Cheyenne Youth Services Center, Montana. 51 Table A.1 : Youth population by year, age, gender and Indigenous status ( 000s) 83 Table A.2 : Ratio of distinct youths apprehended to population by year, age, gender and Indigenous status (%) 84 Table A.3 : Apprehensions by year, age, gender and Indigenous status 85 Table A.4 : Average daily population sentenced to detention by year, age, gender and Indigenous status, base case 86 Table A.5 : Average daily population remanded in custody by year, age, gender and Indigenous status, base case 87 Table A.6 : Average daily population sentenced to community supervision by year, age, gender and Indigenous status, base case 88 Table A.7 : Average daily population in unsentenced community-based supervision, age, gender and Indigenous status, base case 89 iv

6 Glossary ABS ACT AIHW BCR CBA JDAI MST NSW NT QLD SA SCRGSP TAS UK US VIC WA Australian Bureau of Statistics Australian Capital Territory Australian Institute of Health and Welfare benefit to cost ratio cost benefit analysis Juvenile Detention Alternatives Initiative multi-systematic treatment New South Wales Northern Territory Queensland South Australia Steering Committee for the Review of Government Service Provision Tasmania United Kingdom United States Victoria Western Australia v

7 Key terms Apprehensions Bail accommodation Cautioning Detention Diversion Minor offence Recidivism Reoffend Restorative detention Sentenced community-based supervision Sentenced detention Serious offence Unsentenced community-based supervision Unsentenced detention Warning Apprehensions were defined to be youths who are apprehended and proceed to court for this report due to the intended target population and availability of data. A pre-sentencing accommodation option to ensure youths who have committed crimes too serious to warrant diversion attend court and for this report it is proposed to include a secure element. There are two forms of cautioning formal and informal. Informal cautioning is where a police officer will give a person a verbal or written warning. Formal cautioning generally involves some conditions, such as requiring the young person to attend support services. Detention refers to a young person being either detained on remand or serving a sentence in a detention facility. Cautioning, conferencing, and programs aimed at diverting young people away from the court system and thus assist in avoiding a criminal record. Any offence for which diversion is permitted under Section 39 of the Youth Justice Act (NT). Where a person has multiple episodes of incarceration. Where a person has multiple episodes of apprehension. Supervised living arrangements that take a more holistic approach to a young person s needs, avoiding the negative social and life consequences of being incarcerated. Includes orders such as probation, suspended detention, and parole or supervised release. When a young person is convicted and receives a legal order to serve a period of detention. Unless specified otherwise, an offence for which diversion is not permitted under Section 39 of the Youth Justice Act (NT). Includes orders such as supervised or conditional bail and home-detention bail. Where young people have been detained while waiting for a court outcome, which in the NT is mostly youths who are remanded in custody. Includes any verbal or written warning. vi

8 Executive summary The incarceration and treatment of young people in detention has become a topic of national public concern. In response to alarming reports regarding not only the treatment but also future life outcomes of young people in detention, the Royal Commission into the Protection and Detention of Children in the Northern Territory (the Commission ) was established. The Commission seeks to investigate the treatment of young people in the Northern Territory (NT) justice system, and identify associated deficiencies. Key issues being uncovered by the Commission s recent work call for a new approach to juvenile detention and sentencing, not only in the NT but also globally. Accordingly, a three-pronged response specifically tailored to the NT has been proposed. This new system would employ additional diversionary efforts to keep young people from entering detention in the first place, while also employing a more rehabilitation-focused approach to addressing the needs of those young people who still get detained from re-entering detention. This report outlines the results of a cost benefit analysis (CBA) of implementing these options using a PICO approach i.e. target population, intervention, comparator and outcomes. Intervention and comparator This report discusses the viability and benefits of one holistic intervention compared to a do nothing scenario i.e. diversion and supervision of youths in the NT continues as it has for the last 10 years where funding changes are only to maintain existing levels of supervision and diversion. Thus, the target population for the intervention was young people living in the NT who are aged years old and who are apprehended and would proceed to court for a crime between the 2017 and 2027 financial years. For the purposes of this report, all results are presented for financial years (i.e refers to the financial year). The intervention comprises three components, which act at different stages of the youth justice system, including: 1. diversion: cautioning, conferencing and programs aimed at diverting young people away from the court system and thus assist in avoiding a criminal record; 2. bail accommodation: a pre-sentencing accommodation option to ensure youths who have committed serious crimes (crimes too serious to warrant diversion) attend court, which includes a secure component as a replacement for existing remand and a non-secure component for bail support; and 3. restorative detention: supervised living arrangements that take a more holistic approach to a young person s needs, avoiding the negative social and life consequences of being incarcerated. As noted, the comparator was a do nothing scenario where the diversion and supervision of youths in the NT continues as it has for the last 10 years based on historical trajectories. This scenario is referred to as the base case. The base case model outlines the expected costs to the NT if the youth justice system continues on its current trajectory from 2017 to The model uses established trends in the number of youths who are supervised in various settings (either community or detention) during or after sentencing (either sentenced or unsentenced) 1 relative to underlying demographics to project the expected number of youths who are coming into contact with the youth justice system in the NT. 1 Pathways for youth justice modelled in this report have been based on availability of published data. For example, the AIHW (2017) uses the terms as follows: unsentenced detention where young people have been detained while waiting for a court outcome, which in the NT is mostly youths who are remanded in custody; unsentenced community-based supervision includes orders such as supervised or conditional bail and homedetention bail; sentenced detention when a young person is convicted and received a legal order to serve a period of detention; and vii

9 The total costs of the youth justice system (not including court and police costs) were estimated to be $37.3 million in The daily average costs of the youth justice system are expected to grow rapidly over the coming decade, at an average rate of 6.9% per annum, which is in line with the real growth in youth justice system costs across Australia (Productivity Commission, 2017) reaching $113.4 million by (real dollars). The main driver of this growth is growth in daily detention costs using existing facilities. The main outcomes analysed in this report were: reoffending rates for the 12 months following participation in either a diversion program, or bail accommodation services; and recidivism rates (return to sentenced detention) for the 12 months following participation in restorative detention. Benefits included the change in costs (the interventions were estimated to have a lower per unit cost than the comparator) in the current year and the subsequent pathways for adult incarcerations and future crime costs. Table i shows a list of key modelling inputs and assumptions and how they differ between the base case and the intervention. sentenced community-based supervision includes orders such as probation, suspended detention, and parole or supervised release. viii

10 Table i: Modelling inputs and assumptions Assumption or input Base case Intervention Diversion Diversion participants as a proportion of apprehensions proceeding to court 34% 38% Average cost of expanded diversion programs - $12,053 Reoffending rates (apprehended) within 12 months for those proceeding directly to court (base case) versus those sent to diversion programs (intervention) 46.5% 32.0% Bail accommodation Participants as a proportion of apprehensions proceeding to court - 7.6% Upfront capital costs, $ million Average daily cost of detention (remand vs bail accommodation) in 2017, $ Average daily cost of detention (remand vs bail accommodation) in 2027, $ (real dollars) 1, ,256 1,534 Restorative detention Average daily cost of detention per detainee in 2017, $ 1, Average daily cost of detention per detainee in 2027, $ (real dollars) 3,256 1,312 Upfront capital costs, $ million Recidivism rates (return to sentenced detention) within 12 months 75% 46% Recidivism rates for youth who progress to adult detention 61% 38% Source: Deloitte Access Economics estimates. Intervention 1 diversion The benefits of diversion were measured in terms of reducing the number of people in unsentenced community or detention supervision in the current year, reducing the number who are sentenced to community supervision, and reducing reoffending and therefore apprehensions 2 in the following year (with subsequent benefits for reduced sentences to detention). It was estimated that up to 50% of all apprehensions may be diverted in a given year based on NT Government data (Yick, 2016). 3 The incremental number of diversion participants was modelled as 4% of total apprehensions in any given year, which was based on NSW diversion rates applied to the profile of offences committed by young people in detention or remand in the NT. Therefore approximately 100 apprehensions would be diverted in a given year between 2017 and Thus, given that total diversions were approximately 34% in 2016, the diversion ratio with this component of the intervention would represent 38% of apprehensions. Overall, it was estimated that benefits of increased diversion (without bail accommodation) relative to the base case between 2017 and 2027 would increase from approximately $4.2 million to $6.9 million (in real 2 In this report, apprehensions typically refer to youths who are apprehended and would otherwise proceed to court to align with the target population for expanded diversion. Apprehensions were based on the available data from Yick (2016). 3 Homicide and related offences, dangerous and negligent acts endangering persons, and prohibited and regulated weapons and explosives were all considered to be offences that would still require court processes. ix

11 discounted dollars), respectively (as shown in Table ii). Over the same period, costs are expected to increase from $1.0 million to $1.5 million (in real discounted dollars). The total net benefits of increased diversion were estimated to be $52.2 million. Table ii: Overall costs and benefits of diversion intervention, real discounted dollars (rate = 7%) Cost or benefit ($m) Benefits of diversion programs Costs of diversion programs Net benefit Source: Deloitte Access Economics calculations. Note: results are for financial years where 2017 refers to , and so on for other years. Components may not sum to totals due to rounding. Intervention 2 bail accommodation The benefits of bail accommodation relative to the base case were measured in terms of reducing the number of people in unsentenced detention, and reducing the number of people sentenced to detention for a breach of justice order. Bail accommodation would be delivered to approximately 7.6% of total apprehensions following diversion, with considerably lower unit costs than current remand detention approximately $797 per youth per day compared with $1,605 per youth per day. Overall, it was estimated that benefits between 2017 and 2027 would increase from approximately $20.7 million to $31.6 million (in real discounted dollars). Over the same period, operational costs are expected to increase from $7.2 million to $11.1 million (in real discounted dollars), and there would be associated upfront capital costs of $15.3 million a total of $22.5 million in The total net benefits of bail accommodation were estimated to be $183.3 million (in the absence of diversion and restorative detention). Table iii: Overall costs and benefits of bail accommodation, real discounted dollars (rate = 7%) Cost or benefit ($m) Benefits of bail accommodation Costs of bail accommodation Net benefit or cost Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. Intervention 3 restorative detention The primary financial benefits of restorative detention are linked to reduced reoffending, with fewer young people returning to sentenced detention, although there are also lower associated costs for youths who remain in sentenced detention. 5 As with diversion and bail accommodation, there are also fewer crimes and adults sentenced to detention who have a juvenile record. 4 Future benefits and costs were discounted using an annual real discount rate of 7% per annum, which is a standard rate recommended by the Department of the Prime Minister and Cabinet (2016). Discounting future costs and benefits is necessary due to time-preference where society would rather have a dollar today than in the future and to reflect the opportunity cost of money i.e. money spent today could be diverted to alternate uses and earn a rate of return. Thus discounting savings in the future allows for comparability with any costs incurred today. 5 From a societal perspective, enabling youth to get on with education and work would bring greater benefits in the long run, but data limitations do not permit these benefits to be modelled here. x

12 After diversion and bail accommodation, 66 youths were expected to be sentenced to detention in 2017, of which 15 would be newly sentenced and 52 would be returning from a prior incarceration (noting there were 69 youths sentenced to detention in 2016 and the recidivism rate was estimated to be 75%). Table iv: Unique youths sentenced to detention after diversion and bail accommodation, no restorative detention Total sentenced Returning to youth detention (r=75%) New sentences Exit youth justice as youth (r=25%)* Exit youth justice as adult (r=32%) Returning to adult detention (r=61.5%) Net exits from detention Source: Deloitte Access Economics calculations. Note: results are for financial years. Total exits calculated as the sum of the number of youths who exit youth justice as a youth plus those who become an adult and do not return to adult detention. * calculated after youth exit as an adult. r represents the relevant rate or proportion e.g. for return to youth detention it is the recidivism rate, while for exit youth justice as adult, it represents the proportion of youths who become an adult in each year. With restorative detention, 66 youths were still expected to be sentenced to detention in However, the pathways from there are substantially lower due to decreased recidivism rates (approximately 40% lower). In chapter 3, the upfront capital costs to implement restorative detention were estimated to be $17.4 million (in the presence of bail accommodation and greater diversion), which would result in total benefits of approximately $159.7 million (real dollars, discounted at 7%). The benefits are likely to far outweigh the capital costs of building the new facilities. The benefits were estimated to increase from approximately $8.6 million to $16.4 million between 2017 and 2027 (in real discounted dollars). The total net benefits of restorative detention were estimated to be $142.3 million (in the absence of diversion and restorative detention). Table v presents the summarised results for restorative detention. Table v: Overall costs and benefits of restorative detention, real discounted dollars (rate = 7%) Cost or benefit ($m) Costs of restorative detention* Benefits of restorative detention Net benefit or cost Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. * operating costs of restorative detention have been included within benefits (i.e. the cost differential is a benefit). Costs have been modelled in the presence of diversion and bail accommodation for simplicity. xi

13 $ millions Northern Territory youth justice models Cost benefit analysis results Combined, the modelled interventions would result in total discounted benefits of $473.2 million (discounted, real dollars) between 2017 and The total cost of the intervention was estimated to be $137.7 million (discounted, real dollars) between 2017 and The per year discounted costs, benefits and net benefit/cost is shown in Table vi. Table vi: Overall costs and benefits of intervention, real discounted dollars (rate = 7%) Cost or benefit ($m) Costs of intervention (discounted, r = 7%) Benefits of intervention (discounted, r = 7%) Net benefit or cost Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. * operating costs of restorative detention have been included within benefits (i.e. the cost differential is a benefit). The total net benefits were estimated to be $335.5 million, and the benefit to cost ratio (BCR) was estimated to be 3.4 meaning for every $1 the NT Government spends on the interventions, the return to the NT community would likely be $3.40 on average. This estimate is conservative as it does not include subsequent benefits for productivity and wellbeing which can be substantial for adult interventions (Deloitte Access Economics, 2013). However, there are a number of limitations to the data outlined in the background to the report. The net benefits derived from the intervention can be separated into three distinct components: cost of crime avoided, adult outcomes and changed in sentenced numbers due to increased diversion. Change in the number of young people now avoiding a sentence because of diversion programs is the largest contributor to the growing net benefits across the ten year time horizon (as shown in Chart i). Chart i: Breakdown of total benefits following intervention Reduced supervision costs Avoided adult incarceration costs Avoided crime costs Source: Deloitte Access Economics calculations. Sensitivity analysis One-way sensitivity testing revealed the BCR of the proposed intervention was likely to range between 2.8 and 4.2 depending on some of the key assumptions made. The results were most sensitive to the costs of bail xii

14 accommodation and the expected costs of detention into the future. The net benefits of the proposed intervention are likely to range between $234.3 million and $430.4 million. Chart ii shows the results of the one-way sensitivity analysis and expected ranges. Chart ii: Results of sensitivity analysis, BCR and net benefit, real dollars, $ millions Default values BCR 3.4 Detention costs grow at 3.0% instead of 6.9% BCR 2.8 Additional 12% diversion instead of 4% Additional 20% diversion instead of 4% Average cost of diversion programs $20,342 instead of $12,053 in 2017 Reoffending rates for expanded diversion programs 40% instead of 32% Average daily cost of bail accommodation $996 instead of $797 in 2017 Average daily cost of restorative detention $1,363 instead of $682 in 2017 Construction cost of restorative detention $1m instead of $630K per bed Recidivism rates of restorative detention 25.0% for youths, 22.6% for youths becoming adults Recidivism rates of restorative detention 65.0% for youths, 59.7% for youths becoming adults Discount rate of 3% Discount rate of 10% BCR 3.8 BCR 4.2 BCR 3.2 BCR 3.4 BCR 2.9 BCR 3.2 BCR 3.2 BCR 3.6 BCR 3.1 BCR 3.6 BCR Net benefits (real $ millions) Source: Deloitte Access Economics calculations. A multivariate sensitivity analysis was conducted to identify the likely worst case given the available range of evidence presented throughout the report, which included changing the expected growth rate, cost of diversion programs, reoffending and recidivism rates for diversion and restorative detention, and changes in the construction costs and average daily cost of bail accommodation and restorative detention. When considering the combination of all of these measures, the BCR was estimated to be 1.4 and the discounted net benefits were estimated to be $79.6 million. The total discounted benefits were estimated to be $259.1 million and the total discounted costs were estimated to be $179.4 million. Conclusion Because in many cases, evidence has had to be drawn from outside the NT, conservative parameters have been employed wherever there was a range to choose from. For example, the smallest observed impact of restorative detention on recidivism rates has been used, new fixed capital costs have been treated as an upfront lump sum rather than being amortised and discounted over future years, and costs of detention have been assumed to grow at considerably less than historical rates. Any relaxation of these assumptions would only further strengthen the existing large net benefits. Conversely, even in the multivariate analysis where the set of parameters were set to the worst case, the BCR still favoured intervention over the base case (BCR of 1.4 to 1). The available evidence strongly supports reforms to increase the use of diversion and bail accommodation, and there is a compelling case that restorative detention substantially reduces recidivism rates. xiii

15 Thus, implementation of interventions by the NT Government for youth justice would deliver real benefits and help create life changing opportunities for young people at risk of entering the justice system. By activating interventions that work to address the root cause of offending in young people in a holistic manner, these young people would be given the opportunity live a full and productive life in the community. Deloitte Access Economics xiv

16 Number per 10,000 Northern Territory youth justice models 1 Background Deloitte Access Economics was engaged by the Royal Commission into the Protection and Detention of Children in the Northern Territory (the Commission ) to undertake a cost benefit analysis (CBA) of an alternative youth justice model in the Northern Territory (NT). The rate of detention 6 of young people in the NT is substantially higher than any other state or territory in Australia. While other jurisdictions have seen improvements in young people s detention rates across Australia over the last 5 years, the NT has experienced an increase in detention rates over the same period from 55 to 63 per 10,000 young people. Chart 1.1: Young people aged in detention on an average day, by legal status and jurisdiction, (rate) NSW Vic Qld WA SA Tas ACT NT Aust Sentenced Unsentenced Source: Australian Institute of Health and Welfare (AIHW) (2017). The number of young people apprehended in the NT each year has grown substantially over the last decade increasing by 36% from 2006 to 2016, with Indigenous males comprising the greatest share of that growth (Royal Commission into the Protection and Detention of Children in the Northern Territory, 2017). Young people are being apprehended from a younger age, with an increase in the number of year olds in detention (growing 168% between 2006 and 2016). A total of 94% of young people in detention in the NT during were Indigenous (Yick, 2016). The peak age for offending in the NT, as with most jurisdictions, is late teens to early twenties (ABS 7, 2017a), which is also applicable to young males, who constitute over 80% of young people in detention in the NT. Research has shown that incarcerating youths can interrupt and delay the normal pattern of ageing out of the system since detention disrupts their engagement with families, school and work (Holman and Ziedenberg, 2006). Accordingly, this report outlines a three-pronged response to reduce the number of young people in detention in the NT by preventing young people from entering detention in the first place and rehabilitating young people who will still go to detention. The modelled intervention includes an increase in diversion, bail accommodation and detention that is restorative, and rehabilitates young people in detention ( restorative detention ). The interventions act at different stages of the youth justice system. These stages are termed pre-court diversion (defined to include bail accommodation) and post-sentence. The bail 6 Detention refers to a young person being either detained or serving a sentence in a detention facility. 7 Australian Bureau of Statistics. 15

17 accommodation aspect is an alternative to unsentenced detention (remand) and helps ensure youths meet bail requirements and turn up to court. Some key terms in this report are outlined as follows. Youths are young people aged between 10 and 17 years of age, who have committed, or are at risk of committing, an offence. Youths that come into contact with the justice system may be diverted, involving the delivery of verbal or written warnings by police, or formalised programs aimed at targeting the root causes of the offending behaviour. If an offence is serious enough, a young person will formally enter the justice system at the court level. Prior to trial, there are a number of options for young people awaiting their judgement. Young people may be detained while awaiting a court outcome, which includes either: unsentenced detention, including remand in custody; or remain unsupervised in the community. Following presentation at court, a young person may be ordered to: unsentenced community-based supervision, which includes supervised or conditional bail, and home detention bail; sentenced detention, involving a legal order to serve a period of detention; or sentenced community based supervision, including orders such as probation, suspended detention, and parole or supervised release. 1.1 Overview of modelled intervention This report outlines a three-pronged response which may assist to ameliorate the current rates of supervision in the NT. The intervention places greater emphasis on diversion, the use of bail accommodation (both secure and non-secure) to serve as an intermediary while young people are awaiting a court finalisation, and an alternative form of detention that focuses on rehabilitation and restoration of young people in detention restorative detention. In the base case, the comparator was considered to be a do nothing scenario where the diversion and supervision of young people in NT continues on the trajectories based on the past 10 years. Funding changes are only to maintain existing (or increase in line with demographic trajectories) service provision for current programs and services. The target population for the intervention was considered to be young people living in the NT who are aged years old and who are apprehended for a crime between the 2017 and 2027 financial years. The following sections outline the proposed models and the context underlying each model Diversion Diversion aims to prevent further offending by minimising a person s contact with and progression through the criminal justice system and reduce the impact of the criminal justice system. Where necessary, diversion can include appropriate interventions with young people that have come into contact with the justice system and can help to avoid negative labelling and stigma associated with criminal conduct (Joudo, 2008) Background and overview of diversion The NT offers a range of diversionary programs to youths at risk of, and already in contact with, the juvenile justice system. Young people who have committed an offence may receive a verbal or written warning or be placed in a diversion program (including conferencing with the youth and victim or family), which will generally occur before proceeding to court. 8 Diversion programs may be specifically tailored to the needs of young people. For example, the NT currently offers drug and alcohol programs, camp programs where young people undertake various activities such as hiking, and various community programs where young people generally work with not for profit organisations. 8 Diversion can also occur after proceeding to court, although data provided by the NT Government to the Royal Commission (personal communication, June 2017) indicate only 7% of referrals for diversion programs are from a court source. 16

18 In the Budget, the NT Government committed $18.28 million to improve the youth justice diversion system. This was comprised of (NT Government, 2017): $8.05 million to support youth outreach and re-engagement; $4.95 million for youth diversion including intensive camp programs for young people; $4.86 million for bail support programs; and $0.42 million to establish a legislative framework for the development of a contemporary youth justice and rehabilitation system, including an independent oversight mechanism Diversion intervention For this report, it was assumed that eligibility for diversion would need to be altered (or sufficient efforts would need to be made to utilise alternative diversion programs) to enable more youths to participate. Specifically, it was assumed that a youth could be eligible for diversion programs more than twice and that diversion would be based on the NT specific crime profile, but with greater rates of diversion (based on observed experience in NSW). Thus, the proposed diversion model would in essence be based on seriousness of offence, and still recognises that some offences will not be diverted due to public safety considerations. Specifically, the increase in diversion programs was assumed to be equally allocated across a range of programs including: additional conferencing; greater use of youth camps; and increased use of education, alcohol and drug programs and other targeted interventions. There has been a limited number of evaluations performed on NT programs, and as such, evidence for effectiveness and inputs necessary for modelling the benefits of (expanded) diversion requires assumptions. The evidence used as modelling inputs is presented in Chapter Bail accommodation Bail accommodation is a proposed intervention that is pre-sentencing (rather than pre-court) diversion. The primary purpose of bail accommodation is to ensure youths who have been released on bail attend court, and to provide a place for children or young people on bail who have no other alternative address to reside. For the purposes of the intervention, it was also envisaged that there would be a secure version of bail accommodation that could serve as a replacement for remand while young people await their court sentencing outcome Background and overview of bail accommodation A number of youths who are given bail do not then appear in court. NT Government (Yick, 2016) data show that, behind burglary and assault, breach of justice orders represent the largest single category of youths in sentenced detention. Penalties for breach of justice orders are substantial the same data show that the average sentence length is nearly four months (115 days). 9 Some young people who end up in remand are there as they may have no alternative address to reside, and it may not reflect the crime they allegedly committed. PWC (2017) suggest this is one of the biggest barriers to releasing Indigenous young people on bail. As such, a bail accommodation program would reduce the number of Indigenous young people in detention while on remand. There are currently only two pre-trial pathways open to NT youths charged with offences serious enough to warrant a court appearance: remand or bail. 10 Under the Youth Justice Act, detention (including unsentenced detention) should be a last resort for children and young people. Yet evidently those given bail instead often do not front the court on the due day Yick (2016) shows that in almost half (47%) of NT youths given bail subsequently committed breach of bail offenses and thus may receive a subsequent sentence. Bail support programs exist in various forms for all other states and territories, but not for the NT (although bail support initiatives were announced in the NT Budget). For the NT, a Bail Accountability Program 9 Sentence length does not necessarily align with time spent in supervision. 10 Technically, they are remanded on bail, or remanded in custody, but the terms bail and remand (respectively) are commonly used. 17

19 is being developed by Territory Families, to be rolled out in Darwin and Alice Springs. By providing supported accommodation and supervision to young people awaiting trial, young people who do not otherwise have a place to live, or require additional support will be assisted. Overall, this is intended to help reduce the number of young people in detention (Kerr, 2017) Bail accommodation intervention For this report, it was assumed that the full Bail Accountability Program would be initiated. Bail accommodation support would be available to young people who require a safe place to reside while on bail. The program is also intended to offer (Kerr, 2017): food, shelter, education, health and wellbeing services, dependant on the young person s needs; access to support services that will assist their youth justice concerns; electronic monitoring infrastructure including police bail electronic monitoring; transport and other transportation-related services required; supervision to ensure the safety of young people living in the facility; and a Bail Support Advisory Service to both police and young people. Specifically, bail accommodation was modelled to reduce the number of young people who are remanded in custody. Based on limited evidence from the United Kingdom (UK) (Freeman, 2008), the analysis assumed that two-thirds of young people on remand (excluding those who would be diverted through expanded camps and other programs) would receive bail accommodation services, while the remaining one-third would require a secure bail accommodation facility. The secure accommodation facility was not included within the proposed model by Territory Families, and while not strictly bail accommodation, a secure model was envisaged to fully replace existing remand in current detention facilities Restorative detention A secure accommodation model will still be required to hold children who cannot be diverted, with at least a facility in Darwin and a facility in Alice Springs. However, the facilities would be new purpose built, small scale, home style facilities in community settings. Staff would be trained youth specialists focused on delivering interventions to young people rather than simply being there for punitive measures. It was envisaged that this model would help reduce future contact with the juvenile justice system, improve youth safety, and increase the likelihood of young people in detention returning to school or work Background and overview of restorative detention Restorative detention models have been used to successfully alter youth justice outcomes in some states within the United States (US). Restorative detention facilities offer high quality facilities and programs to young people along with a short amount of time spent in detention. Missouri has created humane, constructive and positive youth programs that have produced lower rates of recidivism, a good safety record, and positive youth outcomes with a smaller budget than a number of other states, partly driven by shorter lengths of stay. In Missouri, youths who do not have an extensive offending history and who are at a lower risk of reoffending usually stay in a group home for 4 to 6 months. Those who stay in moderately secure facilities spend roughly 6 to 9 months in a care facility. Secure care facilities still have a length of stay of around 9 to 12 months (Mendel, 2010). While not strictly a model of restorative detention, the Juvenile Detention Alternatives Initiative (JDAI) has reduced reliance on pre-trial detention for youths awaiting placement due to reducing the size of detention facilities and populations and focusing on a model more closely aligned with strategies to reduce youth detention numbers (Mendel, 2014). Restorative detention systems have required substantial improvement in staff training and education, and a change in the mix of staff employed. In the early 1980s, Missouri made the decision to re-define the role of its staff. Staff were no longer working in their traditional role as guards or correction officers, but instead were working as youth specialists and were responsible for the safety and care of the youth. 18

20 Restorative detention intervention There are a range of approaches to restorative detention facilities. Arifuku et al (2010) provide a useful overview of the characteristics of a model that rehabilitates young people in detention, and the proposed restorative detention modelled for this report. Key features include (Arifuku et al, 2010): an open facility with minimal security controls rather than a heavily secure locked facility; a continuum of services rather than limited services; strong relationships between staff and young people in detention, where staff are often counsellors, teachers and others trained in providing effective therapeutic interventions; a treatment model that incorporates family and community rather than limiting interactions; greater structure around programs, rules and regulations and a focus on facilitating exploration of rules and boundaries with the aim of youths internalising and understanding those boundaries; and the model results in cognitive change and development rather than improved behaviour compliance. Specifically, the modelled restorative detention intervention was based on Jara (2013), where the model of care is an intensive therapeutic intervention in secure but small facilities. The facility was based on the New Beginnings Youth Development Center in Washington, DC, which includes: accommodation buildings; an education building; a gym; and a warehouse. For modelling purposes, it was assumed that restorative detention would only be provided to young people who are sentenced to detention, while young people who are remanded in detention would receive alternative services (through a secure bail accommodation model above, which does not necessarily focus on rehabilitation and therapeutic intervention given the shorter time frames than sentenced detention) Summary of interventions Each intervention diversion, bail accommodation and restorative detention could work together to improve outcomes for youths who are supervised in the NT. The proposed interventions will act at different stages and for different young people in contact with the justice system. More youths would be expected to be diverted to other, appropriate diversion programs, such as driver education courses, camp programs, and alcohol and drug programs. For youths who are likely to not meet bail conditions (or have already failed to meet bail conditions), they would be expected to receive bail accommodation support services either secure (to replace existing remand detention in the NT), or not secure in a support role. Finally, for young people who are still sentenced to detention, they would receive services which would help rehabilitate them and return them to society. Figure 1.1 provides an overview of how the interventions would interact with and alter the existing pathways for young people who come into contact with the youth justice system in the NT. It is noted the diagram is heavily simplified for illustrative purposes only, and there would be numerous legislative and other considerations which are beyond the scope of this report. 19

21 Figure 1.1: Summary of modelled interventions in the justice system Source: Deloitte Access Economics. 1.2 Key assumptions and caveats This report outlines a (mostly) conservative approach to estimate the benefits of the proposed alternative youth justice model for the NT, although a number of assumptions and caveats should be made clear relating to what information is available and what assumptions were necessary to model the costs and benefits of the justice system in the NT. For this report, information was requested from the NT Government to inform a range of modelling inputs. The items requested, their use and any further assumptions that were made are outlined in Table 1.1. Table 1.2: Data requested for modelling Information requested Number of young people in detention by legal status, including characteristics such as age, gender and Indigenous status Average length of stay for young people in detention Compensation claims by staff Data on any deaths and hospitalisation of inmates Comments Data were provided, however a decision was made to use data published by the AIHW. Data were provided, however a decision was made to use data published by the AIHW. Data were provided, although the expected quantum of costs was considered to be immaterial to the benefits reported. Data were provided, however it was not possible to determine whether these were for usual care purposes such as dental services or general ill health. As such, no 20

22 Information requested Comments benefits of a change in assaults on other inmates were included in this report. 11 Costs of running detention facilities, including a breakdown by capital costs, staff costs, staff numbers Any budget allocation for new regular detention facilities Cost data and number of participants in current and past diversion programs, and details about the nature of programs Recidivism rates for young people in detention Data on the offence profile, and pathways through the youth justice system Average daily costs of running detention facilities was provided. Partially provided, although the advice given indicated that no firm cost allocation had been made at the time of writing. Consequently, costs of new regular facilities were conservatively excluded from the modelling. Data were provided, and combined with inputs sourced from the Report on Government Services. No NT specific data were provided for existing programs. Rates were derived from youth justice in other states and triangulated with adult recidivism rates. Data were provided, although a decision was made to use published statistics from the AIHW for young people in the NT as it represented actual outcomes rather than sentenced outcomes. Source: Deloitte Access Economics. The data requested were combined with a range of inputs from official government sources, and a number of assumptions were necessary to estimate the benefits of the proposed model in the absence of data specific to the NT. Assumptions were also required in some instances when data were either incomplete or not fit-forpurpose. It is noted that there are limitations to this approach. For example, as there are no published recidivism rates for young people exiting detention in the NT, it was assumed that the recidivism rates would be equivalent to that of other states and territories in Australia. Another key assumption is that the benefits observed from US models of restorative detention would be realised in Australia, although the applicability of such models in the NT has not been tested and is subject to large caveats. Table 1.2 outlines a list of primary assumptions, and attempts to highlight the limitations and shortcomings to the estimates presented in this report (noting the myriad of ways the NT is different from other jurisdictions in Australia). Table 1.3: Key assumptions and limitations used in the modelling Assumption/ limitation Base case and intervention recidivism rates for detention Benefits of diversion for reoffending rates Comments In the absence of NT specific data, recidivism rates from other jurisdictions were used in the base case, and recidivism rates for the intervention were based on similar restorative detention models in the US. There has been a limited number of evaluations performed on NT programs, and as such, the reduction in reoffending due to diversion was estimated by calculating the average change for diversion programs tailored to the current demographic profile of young people who are apprehended in the NT using the results published by Cunningham (2007). These rates are potentially outdated, although the study does include a comparison group which was not available in more recent annual reports published by the NT Youth Diversion Unit (Renfree, 2017). 11 Mendel (2010) reports that under restorative detention in the US (compared to regular detention), youth-on-youth assaults are 4.5 times lower, youth-on-staff assaults are 14 times lower, sexual assaults are 20 times lower, and there have been no suicides for over 25 years, compared to more than 100 in regular detention over the same period. 21

23 Assumption/ limitation Cost and mix of expanded diversion programs Eligibility for diversion Cost growth for supervision types Comments Average cost of new programs were based on the average cost per participant for existing diversion programs. Participants were allocated to new programs equally, on the assumption that discretion would be used by the Youth Diversion Unit to appropriately allocate participants to the program that would be effective for each youth. In reality, effectiveness is highly dependent on a range of characteristics for each youth, and mix of expanded programs may not be appropriately allocated, although the cost range is not excessively large. It was assumed that a youth could be eligible for diversion programs more than twice, and that eligibility for diversion would match that of NSW (the state that appears to divert the most young people from the justice system). Eligibility is decided at the discretion of the Youth Diversion Unit for a number of offence types and methods would need to be decided upon to reach the targeted diversion rates. Although costs of supervision, particularly in detention, have been growing rapidly, it was assumed that this would not continue as rapidly in the future. This is because a substantial proportion of costs are likely due to external factors of the current facilities (i.e. deterioration). Therefore, real growth in the detention costs was assumed to be the average real growth for all youth detention in Australia over the last four years. Costs of restorative detention and bail accommodation were assumed to grow at this same rate. Costs of community correction were based on the historical growth in youth community correction supervision for all of Australia over the last four years. Number of days supervised Costs of bail accommodation For youth who receive either bail accommodation services or restorative detention, the average number of days in the supervision types was assumed to be the same as in the base case remand and regular detention, respectively. The cost of bail accommodation is based on the proposed budget for Bail Support Accommodation. This included $15.3 million for construction costs, which would provide 58 beds in total across Darwin, Alice Springs, Tennant Creek and Katherine. Ongoing annual costs are expected to total $16.87 million per year. As such, the ongoing annual cost of bail accommodation was expected to be $290,862 per bed, or $797 per child in bail accommodation per day, which was grown in line with average detention costs. It is possible the cost may be higher or lower depending on whether secure models would be included. Participants for bail accommodation A third of those currently on remand were assumed to remain there, the remaining remandees (excluding those who are diverted in the other interventions) were assumed to receive bail accommodation support. This assumption is based on United Kingdom (UK) data from Freeman (2008). For the third who continue to be remanded in custody, it may be an option to have secure bail accommodation facilities until they receive their court finalisation outcome. Thus, the remaining third were also modelled as receiving bail accommodation (in terms of costs). Overall, remand would be replaced with bail accommodation until court finalisation occurs. Benefits of bail accommodation on sentenced detention outcomes Costs of restorative detention Data from the AIHW (2016) showed that of youths who are first supervised in the community while awaiting court finalisation, 66% will have a subsequent pathway through youth justice compared with 73% of youths in unsentenced detention. The relative reduction was applied to court reoffending rates to estimate the reduction in reoffending rates attributed to bail accommodation. In the absence of NT specific data, costs were based on similar restorative detention models in the US. In the no change base case, it is conservatively assumed that the current Berrimah facilities would be used for another decade, rather than any new regular youth detention facilities being constructed. Rather than median staff costs, it was assumed that staff in restorative detention facilities would be paid in the top 25% of salaries for their professions, due to recruitment and retention issues in the NT, and the need to have more highly trained staff. 22

24 Source: Deloitte Access Economics. In summary, this report outlines the likely or indicative benefits that may result from an alternative youth justice model as described, and it is recognised that key assumptions will greatly influence the expected costs and benefits and this is really a best efforts exercise in the absence of hard data. As such, sensitivity analysis was conducted on key assumptions, and a range of results are presented to ameliorate the influence of any assumptions on the final outcome. 23

25 2 Base case To model the outcomes for the intervention, it is necessary to know what the youth justice system in the NT will look like in the future. A ten year modelling horizon has been used, where the base case outlines the expected costs to the NT if the youth justice system continues on its current trajectory. 2.1 Base case model To model the base case, there are a number of modelling inputs and assumptions. A demographic model of NT children capturing relevant factors such as age, gender and Indigenous status was constructed to inform the base case projections. Primarily the model uses established trends in the supervised population relative to underlying demographics to project the expected number of youths who are: apprehended; supervised in unsentenced community or detention settings; and supervised in sentenced community or detention settings. The expected trajectories by age and gender are outlined in Appendix A, and includes the: number of youths in the NT, including mix by age, gender and Indigenous status which was used to determine apprehension numbers; and average daily number of youths supervised by setting by age, gender and Indigenous status. The following sections discuss factors relevant to the base case, including brief methods, and cost inputs, particularly for the costs of supervision and the existing diversion programs. 2.2 Base case trajectories To estimate the trajectory for the current NT youth system, it was necessary to get population projections by age, gender, and Indigenous status (NT Treasury, 2016). Apprehensions data were obtained from the NT Government, which defined age groups as 10-14, 15-16, and 17 (Yick, 2016). Any apprehensions that occur over 18 years of age were excluded from the modelling. Trends in the Indigenous status, gender and age of youths in the NT between 2016 and 2027 are shown in Chart 2.1, Chart 2.2, and Chart 2.3, respectively. As a share of the population living in the NT, Indigenous youths are expected to essentially be stable (a very marginal decline) as a share of the total population over the coming decade based on population projections by the NT Treasury (Chart 2.1). The changes in the underlying demographics are not sufficient to substantially alter the supervision of children and young people in the youth justice system. Of more importance, there are projected to be more youths aged (Chart 2.3), which will grow as a share of the total youth population meaning that more younger children will be in supervision over time and the average age of youths under supervision will decline. 24

26 Population ('000s) Population ('000s) Northern Territory youth justice models Chart 2.1: Youth population by Indigenous status in the NT, 2007 to Non-Indigenous Indigenous Source: ABS (2016), NT Treasury (2016) and Deloitte Access Economics calculations. Chart 2.2: Youth population by gender in the NT, 2007 to Male Female Source: ABS (2016), NT Treasury (2016) and Deloitte Access Economics calculations. 25

27 Population ('000s) Northern Territory youth justice models Chart 2.3: Youth population by age in the NT, 2007 to Source: ABS (2016), NT Treasury (2016) and Deloitte Access Economics calculations. Historical apprehension patterns were based on data from the NT Government (Yick, 2016). The apprehensions were derived relative to the respective population subgroups. Trends were fitted to the ratios of apprehensions to the total number of youths in each group, which were used to estimate the expected change in apprehensions per youth in the NT over the modelling horizon. Linear trends were used to project ratios that were growing as a share of the population and logarithmic trends were used to project ratios that were declining as a share of the population (to prevent any single ratio from becoming negative over the modelling horizon). The trends indicate that apprehensions will worsen over the decade between 2017 and 2027, in terms of both total numbers and the apprehensions per distinct youth. That is, more NT youths will be apprehended multiple times in a given year, and the frequency of contact will likely increase as well. 12 The trends in apprehensions are shown in Chart An implicit assumption in the modelling is that the sentence time will not increase dramatically, and it will likely need to decrease to enable youths to be apprehended so many times in a given year. This assumption is not inconsistent with historical trends, based on NT Government data (Yick, 2016), which show that the average length of stay for detention has been declining historically. 26

28 Apprehensions per youth Thousands Northern Territory youth justice models Chart 2.4: Youths apprehended and rate of apprehensions per distinct youth in the NT, by year Apprehensions (RHS) Rate of apprehensions per distinct youth (LHS) Distinct youths apprehended (RHS) Source: Deloitte Access Economics calculations. Note: results are for financial years. Based on past trends, it would be expected that more cases heard in court will finalise with an outcome of detention and community orders over time. Very few youths receive sentences for home detention and suspended detention, and there are expected to be fewer monetary and other orders over the coming decade. While there are some trends in court finalisation outcomes, the profile of youth offences does not appear to have relevant trends meaning the seriousness of offence appears to be stable over time. While there are a growing number of crimes in the NT, overall and relative to the population, no substantial changes were projected based on past crime profiles or the trends in court finalisation patterns based on data from Yick (2016). The trends in the offence profile are shown in Chart 2.5. Chart 2.5: Historical and projected offence profile for young people in detention in the NT, total and relative to population % 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% Acts intended to cause injury Breach of justice order Robbery, extortion and related offences Other Unlawful entry with intent, burglary, break and enter Prohibited and regulated weapons and explosive offences Dangerous or negligent driving Offences as a proportion of children and young people (RHS) Source: Yick (2016). There are a number of important realisations from NT court finalisation data (Figure 2.1). 27

29 1. A number of youths are diverted from court processes. 2. Bail occurs, although there are a substantial number of youths who receive a breach of justice order, leading to increased numbers in detention both remanded in custody and sentenced. 3. For every 100 apprehensions, approximately 66 cases are heard in court 13, of which around 16 are withdrawn or acquitted. Approximately 50 of the 100 cases will be found guilty, and 14 will be given a custodial sentence with 13 apprehensions being in sentenced detention. Approximately 36 apprehensions would therefore receive a non-custodial order i.e. sentenced to community including monetary penalties. 4. Of people in detention, approximately 25 are on remand, and 11 are sentenced, noting this is representative of average daily numbers over a number of years (many more youths go to detention each year). Figure 2.1: Average pathways of youths apprehended in the NT, to Source: Yick (2016). However, sentenced outcomes differ from real world numbers of youths who are detained in a given year, or sentenced to a community order. For modelling purposes, the apprehension numbers in a given year were associated with the number of unique youths who go through any of the following pathways: unsentenced detention; unsentenced community-based supervision; sentenced; and sentenced community-based supervision. 13 Based on number of court finalisations, which may not accurately represent the number of apprehensions as there may be multiple court finalisations for one apprehension or vice versa. 28

30 The AIHW s Youth Justice in Australia series of reports provide key characteristics about the number of young people who experience each type of supervision, and also the total number of youth who experience any type of supervision in the NT. 14 Based on the Youth Justice in Australia for (AIHW, 2017): 260 youths, or 13.3%, of apprehensions, were remanded in custody; 52 youths, or 2.7% of apprehensions, were supervised in unsentenced community-based settings; 69 youths, or 3.5% of apprehensions, were sentenced to detention; 245 youths, or 12.5% of apprehensions, were supervised in sentenced community-based settings; and 371 youths, or 19.0% of apprehensions, were supervised in any setting. The ratio of supervised youths to total apprehensions was held constant for each type of supervision over the coming decade, and applied to the total estimated apprehensions to estimate the number of youths who would be supervised in any setting over the coming decade. The growth is shown in Chart 2.6. Given that diversion acts on the number of apprehensions in subsequent years, it is then possible to model the effect of diversion on supervision numbers (holding the ratios constant). Chart 2.6: Projected numbers of supervised youth in the NT Unsentenced community Sentenced community Unsentenced detention Sentenced detention Total supervised Source: AIHW (2017) and Deloitte Access Economics calculations. Note: results are for financial years. It was estimated that the number of supervised youths in the NT will grow by almost 70% (based on growth in apprehensions) from 2016 to 2027, representing 628 youths, or 257 more youths in 2027 than in A number of apprehended youths will experience more than one supervision type in any given year (Table 2.1). A more detailed breakdown of the expected number of supervised youths is included in Appendix A outlining the breakdown by age, gender and Indigenous status. 14 Data relevant to the NT youths of interest were obtained from the latest reports in the series as necessary, to fill in information about historical trends. 29

31 Table 2.1: Supervised youths by year in the NT Supervision type Sentenced detention Sentenced community Unsentenced detention Unsentenced community Any supervision Source: Deloitte Access Economics calculations based on AIHW (2017). Note: results are for financial years Diversion in the NT Historically, funding for youth diversion has been around $5 million annually, including grants to other organisations to provide diversion programs and funding to the NT Youth Diversion Unit to facilitate diversion and provide warnings and cautions (Renfree, 2017). Data from the NT Youth Diversion Unit (Renfree, 2017) provides information on five types of diversion programs as shown in Table 2.2, although there are some additional diversions in other programs such as existing boot camps and residential facilities. Table 2.2 shows the number of youths participating in each program category over the last four years. The majority of these participants were Indigenous youths. More than 669 young people either completed or failed a youth diversion program in , representing approximately 34% of youth apprehensions in the NT for the year. Table 2.2: Number of participants by diversion category, to Program category Drug diversion Family conference Verbal warnings Victim/offender conference Written warnings Total Source: Renfree (2017). There were no apparent trends in the ratio of diversions to apprehensions, so the ratio was held constant over the modelling horizon (at 34%) for the base case. 2.3 Cost inputs Cost inputs used to inform this report include the costs of apprehensions, costs of supervision, costs of adult incarceration, and average length of stay Costs of apprehensions The Australian Institute of Criminology (Smith et al, 2014) estimated that the costs of all crime in Australia in 2011 were $47.5 billion, which included $23.0 billion directly due to crimes, criminal justice costs of $16.3 billion, and other costs of $8.3 billion. The other costs were comprised of victim assistance, security industry, insurance administration, and household precautions. Corrective service costs were estimated to be $3.8 billion, which are netted from the total costs of crime as they are costed separately in the modelling for this report. The incidence of crime in Australia was used to derive average costs by crime types which was calculated to be $3,590 per crime in dollars (Smith et al, 2014). The average costs of various crimes 30

32 were then weighted by the NT crime profile for young people in detention to estimate the costs of crime in the NT 15 as $4,179 per crime. The costs of criminal justice, victim assistance, the security industry, insurance administration and household precautions were assumed to scale in line with the NT youth crime profile representing an additional $4,454 per crime committed in the NT. In the NT, youths were charged with 8,233 offences and there were 1,953 apprehensions, representing an average of 4.21 offences per apprehension. The average cost of crime per apprehension in the NT (net of corrective service costs) was therefore estimated to be $36,391 ($8,633 * 4.21), which was applied to the number of averted apprehensions to estimate the benefits of averted youth crime. The cost breakdown is shown in Table 2.3. Table 2.3: Cost of crime, inputs by cost type Cost type Total Australia ($ million), Per apprehension in the NT ($), Costs of crime to victims 22,995 17,615 Criminal justice 12,419 12,453 Victim assistance 1,877 1,438 Security industry 3,400 2,605 Insurance administration Household precautions 2,360 1,808 Total crime and other costs 43,718 36,430 Assets confiscated* Total 43,668 36,391 Source: Smith et al (2014). Components may not sum to totals due to rounding. * represents money and assets recovered from crime which are sold and may be returned to victims or realised as revenue Costs of supervision Costs of supervision were derived from the Productivity Commission s Report on Government Services and from the NT Government (Campbell, 2016). Costs of supervision, particularly in detention, have been growing rapidly. It was assumed that the costs of detention in the NT were unlikely to continue to rise at the same pace, as a substantial proportion of costs are likely due to external factors of the current facilities (i.e. deterioration). As such, real growth in the detention costs have been constrained to 6.93% per annum over the modelling horizon the average real growth for all youth detention in Australia over the last four years. Costs of restorative detention and bail accommodation were assumed to grow at the same rate. Costs of community correction were based on the historical growth in youth community correction supervision for all of Australia over the last four years an average of 7.14% per annum. A logistic growth curve was estimated based on the existing trends, which accommodated the constrained growth path and matched the total growth over the forecast period. The projected cost series is shown in Chart 2.7. The expected costs of bail accommodation and restorative detention are discussed in sections and 3.3.1, respectively, although they are also shown in Chart 2.7 for comparison purposes. 15 The crime profile for the NT was based on the crime profile for young people in detention (rather than all crime in the NT) as it was the only data available to us at the time of writing the report which reported the primary crime committed. 31

33 Chart 2.7: Costs of detention and community-based supervision per youth per day in the NT (real dollars) 3,500 3,000 2,500 2,000 1,500 1, ,095 2,274 2,441 2,598 2,744 2,883 3,014 3,138 3,256 1,691 1,901 1,534 1,420 1,605 1,293 1,478 1,212 1,150 1,358 1,224 1, ,312 1, ,106 1,264 1,162 1, Detention Community supervision Restorative detention Bail accommodation Source: Productivity Commission (2017), Campbell (2016), Kerr (2017), and Deloitte Access Economics calculations Costs of adult incarceration No direct pathways were modelled for the base case into adulthood due to limitations of available data. However, the effect of diversion on the adult detainee population was modelled incrementally (section 3.3). The costs of adult incarceration were based on results presented in Deloitte Access Economics (2013), where it was estimated the average national costs of an adult Indigenous incarceration (including discounted future recidivism) was $385,915 in Allowing for higher NT incarceration costs 17, but the lower recidivism rates for the 16% of adult NT prisoners who are not Indigenous, adult incarceration costs in the NT were estimated to be $345,256 in The effect of the intervention on the adult cost of incarceration is explored in the intervention chapters Average length of stay The average length of stay by supervision type for the NT was derived from AIHW s (2017) Youth Justice in Australia publication. Historical NT data were obtained for the financial years between 2012 and The length of stay for other years was assumed to be the average of the five year historical period for detention supervision. For community-based supervision, the length of stay was assumed to be the same as for the 2016 financial year. The length of stay inputs are shown in Table 2.4. Length of stay was available by Indigenous status and so it was applied to the expected number of Indigenous and non-indigenous youths in each sentence type. 16 That study also included costs of lost wellbeing from diseases and injuries contracted in prison. As such costs are not included in this study, they have been netted out here. 17 The cost of adult incarceration was based on values from the 2016 Report on Government Services (Productivity Commission, 2017). The costs were assumed to remain constant in real terms, as it has been declining marginally in real terms for the last couple of years. 32

34 Table 2.4: Average length of stay for supervision types in the NT (days) Supervision type Average Indigenous youth Sentenced detention Sentenced community Unsentenced detention Unsentenced community Non-Indigenous youth Sentenced detention Sentenced community Unsentenced detention Unsentenced community Source: Deloitte Access Economics calculations and AIHW (2017). Note: results are for financial years. 2.4 Base case results To estimate the base case costs of youth offending in the NT, it was necessary to know the number of youths who will be supervised in any given year, as well as the average length of stay in supervision for those youths. The historical average length of stay and expected youth entering supervision in each given year were multiplied together to give the total expected days of supervision. This was then converted to the average daily population by dividing by 365. Average daily costs of supervision were applied to the total days spent in supervision. The average daily population and estimated days of supervision are shown in Table

35 Table 2.5: Supervised youth and days of supervision in the NT, by year Supervision type Average daily population Sentenced detention Sentenced community Unsentenced detention Unsentenced community Days of supervision ( 000s) Sentenced detention Sentenced community Unsentenced detention Unsentenced community Total Source: Deloitte Access Economics calculations and AIHW (2017). Note: results are for financial years. Components may not sum to totals due to rounding. Costs of the NT youth justice system (not including court and police costs) are expected to grow rapidly over the coming decade, at an average rate of 6.9% per annum (section 2.3.2), in line with the growth in supervision costs seen for all youth justice across Australia. The total cost was estimated to be $37.3 million in 2017, reaching $113.4 by 2027 (real dollars), which will primarily be driven by growth in daily detention costs using existing facilities. Table 2.6: Costs of supervision in the NT, current year, $ million, dollars Supervision type Sentenced detention Sentenced community Unsentenced detention Unsentenced community Total Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. 34

36 3 Intervention Three interventions have been proposed, which include an increase in diversion programs that are tailored and specific to children and their offences, bail accommodation (both secure and not secure) to reduce the number of children remanded in custody, and restorative detention to rehabilitate young people sentenced to detention. The benefits of each intervention are modelled in slightly different ways. Diversion programs remove children and young people from any pathway in the current year, and reoffending rates at 12 months are used to model the subsequent reduction in apprehensions in the subsequent year. To be conservative about the total benefits of diversion, no subsequent year benefits are included as insufficient evidence was available for the severity and frequency of offending following diversion programs compared to children who go to court. Bail accommodation serves two purposes. First to act as a substitute for remand in the NT, so that all less serious cases of remand receive bail accommodation. Second, it is envisaged that remaining remand cases may need a secure bail accommodation arrangement equivalent to remand, except with higher staff to detainee ratios much like restorative detention. Bail accommodation has benefits of reduced reoffending relative to current remand, and also removes some existing sentenced detention who were on bail and were then sentenced to detention for breach of justice orders. Restorative detention acts to rehabilitate young detainees, reducing the recidivism (return to sentenced detention). There are also associated reductions in apprehension numbers (assumed to be the same as the change in detainee population numbers). All pathways alter the number of apprehensions, and have associated reduced crime costs. There would also be a smaller adult detention population as a result of the interventions. The following chapter outlines the expected savings to the NT community from a societal perspective for the three interventions. The benefits are modelled in a three stage approach, whereby: Diversion acts first: reducing the number of people in unsentenced community or detention supervision in the current year, and also reduces sentenced community (i.e. less serious crimes); and reduces reoffending and therefore apprehensions in the following year. Bail accommodation acts second: further reducing the number of people in unsentenced detention, and also reduces the number of people sentenced to detention for breach of justice order; a secure alternative is considered as an alternative for serious offences in unsentenced detention (in terms of costs); and there are small benefits to reoffending in the following year. Restorative detention acts third: all children sentenced to detention receive restorative detention, which has no first year impacts except for change in costs; and in subsequent years, fewer children reoffend, so fewer return to sentenced detention. Each subsection discusses the costs and benefits of the intervention, and then the subsequent overall net benefits or costs. 3.1 Diversion Based on data from the NT Government (Yick, 2016), it would seem that a number of children who are detained in the NT are there for relatively minor offences. Consequently, it was presumed that a substantial proportion could be diverted to community based programs, which would produce savings to the youth justice system as diversion is cheaper than secure accommodation and it may also result in reduced reoffending rates. The purpose of this section is to outline the savings to the NT community for adopting increased diversion as a suggested reform. 35

37 3.1.1 Costs of diversion intervention The costs associated with increasing government expenditure on diversion programs is inherently founded on the cost per participant for each program category, as discussed below Cost per participant The average cost per person for each program category was derived from existing funding for NT youth diversion in using total funding and the total number of participants. Total funding for the NT Youth Diversion Unit was $4.89 million in , of which agency expenditure was $2.43 million and there were $2.46 million worth of grants provided to other organisations to provide diversion programs. All diversion programs were assumed to have the same agency cost $2.43 million divided by the number of diversion participants (669 in total). Conferencing and drug and alcohol programs were estimated to have additional costs per person of $5,365, which represents grant funding divided by the number of conferencing and drug and alcohol participants (458). The costs of youth camps were based on evidence from Queensland (KPMG, 2015) and incorporated the costs of BushMob s residential treatment program. KPMG (2015) reported average costs per participant for early intervention youth camps in Fraser, Gold Coast and Rockhampton were $13,589, $8,844 and $15,352, respectively. The average cost per participant for a sentenced youth boot camp in Queensland was $44,496 based on an initial target to have 84 participants in the programs (lower end of the range). For BushMob s residential treatment program in the NT, total expenditure on the program was $1.494 million and there were 110 participants during , resulting in an average cost of approximately $13,582 per participant (BushMob Aboriginal Corporation, 2016). Taking a weighted average of the camp programs, it was estimated youth camps would cost $20, The average costs of diversion programs per participant used in the modelling are presented in Table 3.1 (in dollars). Table 3.1: Cost per person, dollars Program category Cost per person Drug diversion $9,290 Family conference $9,290 Victim/offender conference $9,290 Youth camps $20,342 Alternative programs* $12,053 Source: Renfree (2017); KPMG (2015); Deloitte Access Economics calculations. * This represents a weighted average of equal participants in each of the four programs (drug diversion, family conference, victim/offender conference and youth camps) Increase in diversionary expenditure Based on the Steering Committee for the Review of Government Services Provision (SCRGSP, 2016), NSW has the highest rates of diversion out of all states and territories (53.7%). The NSW non-indigenous juvenile diversion rates by crime type were selected to reflect the potential eligibility for diversion for the proposed intervention in the NT. The 2015 NSW juvenile diversion rates by crime type are shown in Table 3.2. It is noted that the estimated eligibility reflects an ideal or aspirational diversion rate for the NT. As noted in Table 1.3, achieving an increase in diversion rates will likely require changes to existing legislation and more flexibility in the types of crimes that can be diverted. Moreover, achieving such eligibility would require an expansion of appropriate and targeted diversion programs for youths, and the effectiveness of any such programs would need to be tested and proven. The potential eligibility for diversion based on NSW by crime type was weighted by the NT crime profile for young people received into detention during Using this approach, it was estimated that at least 38.3% of apprehensions in a given year may be diverted noting the crime types which were not received into 18 Participant numbers for Queensland boot camps were 58, 60, 44 and a targeted value of 84, respectively. 36

38 detention are likely to be less severe. Sensitivity analysis was also conducted using the total diversion rate for NSW and for SA in section 4.4. As a comparison, using NT Government data (Yick, 2016), it was estimated that as many as 50% of total apprehensions could be diverted if all crimes aside from assaults, homicide, abduction and harassment, prohibited and regulated weapons, or robbery, extortion and related offences were diverted. Table 3.2: Juvenile diversion rates and number of apprehensions by crime type, non-indigenous, 2015 Crime type Young people received into detention, NT Potential eligibility for diversion based on NSW diversion rates Homicide and related offences 1 0.0% Acts intended to cause injury % Sexual assault and related offences % Driving under the influence of alcohol or drugs/dangerous or negligent driving % Other Dangerous Acts % Abduction, harassment and other offences against the person % Robbery, extortion and related offences % Unlawful entry with intent, burglary, break and enter % Illegal use of motor vehicle % Theft other % Fraud, deception and related offences % Illicit drug offences % Prohibited and regulated weapons and explosive offences % Property damage and environmental pollution % Public order offences % Traffic and vehicle regulatory offences % Offences against justice procedures, government security and operations % Miscellaneous offences % Total/weighted average % Source: SCRGSP (2016). Costs of diversion were assumed to grow in line with the projected growth for all unsentenced community costs (i.e. real growth of 7.14% as in section 2.3.2). This was applied to the participant numbers to estimate the cost of diversion in the future. 37

39 Table 3.3: Overall costs of diversion, real dollars. Costs Additional participants Cost of programs ($m) Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding Benefits of diversion intervention Effectiveness of diversion in the literature The following subsections outline the benefits of diversion in the literature to provide context, although not all programs were included in the modelling due to data constraints. Youth Group Conferencing Group conferencing is where the youth meets with the victim or victim s legal representative, their family, victim support workers, police, and their legal representative. The aim of the conference is to help the youth to understand the consequences of their actions. An outcome of the conference is to put in place actions that the youth can take to address the underlying reasons for their criminal activity. Conferencing has been found to be effective in reducing offending across Australia and in the NT specifically. Allard et al (2010) reported a 21.1% reduction in re-contact with the juvenile justice system for Indigenous men, a 21.4% reduction for Indigenous women, and for all participants a 24.5% reduction in re-contact (compared to young people who had attended court). KPMG (2010) reported a 23.7% reduction in reoffending (convicted of multiple offences) over the 24 month period following the conference compared to a group of young people who had received a Probation or Youth Supervision Order. Luke and Lind (2002) found a 19.0% decrease in reoffending (convicted of multiple offences) compared to young people who had attended court. Cunningham (2007) reported an 18.0% reduction in reoffending for young people who were diverted in the NT. Youth conferencing is also related to other benefits such as improved educational attainment that, combined with reduced recidivism, improves employment outcomes. Cautioning There are two forms of cautioning formal and informal. Informal cautioning is where a police officer will give a person a verbal or written warning. Formal cautioning generally involves some conditions, such as requiring the young person to attend support services. Due to the lack of monitoring for informal cautioning there are limited data about the efficacy of informal cautioning. Formal cautioning requires follow up and monitoring and as a result, data are available to estimate the efficacy of the program. The effectiveness of formal cautioning varies considerably in the literature 19 : Allard et al (2010) found that formal cautioning of Indigenous men and women reduced re-contact with the juvenile justice system by 0.3% and 8.3%, respectively. The average reduction in recidivism across all respondents (males, females, Indigenous and non-indigenous) was 19.4%. Dennison et al (2006) reported a decrease in re-contact with the juvenile justice system of 11.0%. 19 In each study the reoffending rates of those who had been cautioned were compared to reoffending rates of those who had proceeded to court. 38

40 Indigenous healing programs These programs provide culturally appropriate services to Indigenous youths within the youth justice system. Programs may have a particular focus such as drug and alcohol addiction. The literature search for this report did not uncover any cost-efficacy data for healing programs for Indigenous youth in the NT. As an example, the Canadian Aboriginal Justice Strategy takes into account higher rates of crime, incarceration and victimisation of Indigenous people and subsequently formulates policies, programs and initiatives to address such issues. This strategy has been found to reduce recidivism amongst Indigenous people by up to 13%, although it is not clear if this indicative rate would apply in an Australian setting (Office of Strategic Planning and Performance Management, 2011). Other diversion programs There are a number of other diversion programs, including boot camps, drug courts, and more targeted programs including programs which aim to reduce offending by scaring offenders. Available evidence does not tend to indicate positive benefits, so these programs are mentioned but not modelled. A consistent finding in the literature is that programs that aim to scare the youth off crime, such as prison visits or the scared straight program, can increase the rate of recidivism. Scared straight programs are those that aim to deter future offending by having youths observe prison life and/or meet adult inmates (Petrosino et al, 2013). A study by Drake et al (2009) analysed the impact on reoffending of a scared straight program and found that on average, the program increased recidivism by 6.1%. The Attorney-General s Department (2013) also supported these findings and found that scared straight programs can increase recidivism by up to 26.0%. Similarly, boot camp type programs have also been found to be ineffective at reducing recidivism or increase recidivism. The Attorney-General s Department (2013) found in a review of the literature that boot camp type programs may increase recidivism by up to 10%. Wilderness camps were also found to have no impact on recidivism. For the purposes of this report, no evidence was found which referred to the early intervention youth camps which are proposed by the NT Government. In addition, drug courts such as those offered by the NT Government have not been found to successfully divert young people. Rysavy et al (2011) attribute this to a significant proportion of Indigenous men failing to complete the programs prescribed by drug courts. Summary of effectiveness of diversion programs With regard to the benefits of diversion programs, literature tends to focus on reducing reoffending rates. Table 3.4 summarises the findings from the literature, although NT specific outcomes were reported where available (i.e. Cunningham, 2007). The reoffending rates from Cunningham (2007) were adjusted by a factor of 1.19 to reflect the distributional differences in the cohort of youths who proceed to court relative to all diversions, noting that a larger proportion are Indigenous. The reoffending rates for conference programs would therefore be around 25.1% based on Cunningham (2007) The Northern Territory Police, Fire & Emergency Services annual report publishes reoffending rates for youth conferences in the NT, where the reported rates are substantially lower (15%). The reoffending rates include youths who have commenced but not yet completed diversion, and thus may not appropriately represent final outcomes over a 12 month period. 39

41 Table 3.4: Reoffending rates by diversion program Type of diversion Reoffend within 12 months Source Written warning 22.7% Cunningham (2007) Verbal warning 22.7% Cunningham (2007) Drug diversion 32.0% Office of Crime Statistics and Research (2012) Family conference 25.1% Cunningham (2007) Victim/offender conference 25.1% Cunningham (2007) Source: as itemised in table. Programs were found to reduce recidivism by engaging young people in addressing the behaviours and root causes of their offending behaviour. A meta-analysis conducted by Wilson and Hoge (2012) reported that both caution and intervention diversion programs are significantly more effective in reducing recidivism in young people than incarceration. The authors identify a common theme in the youth diversion discussion, regarding the increased risk of reoffending when a young person is participating in programs with negative peers and values. However, the analysis found that considering the risk level of participants and matching this to the programs they are enrolled in can effectively mitigate the risk, based on limited available evidence. Programs were found to be effective across all offender risk categories. In fact, the greatest reductions in recidivism were programs targeted at medium and high risk offenders (Wilson and Hoge, 2012). While there was considerable evidence for reoffending rates, the literature does not sufficiently focus on seriousness of offence for inclusion in the modelling (without access to more detailed data). For example, Piggott (2015) found that subsequent offence was generally more serious than the initial offence for both caution and conference groups, but did not have data available to allow comparison with matched groups who attend court. In the base case, reoffending rates for youths who would proceed to court were estimated to be 46.5% based on Cunningham (2007). Cunningham (2007) found that 39% of youths who proceed to court reoffend within 12 months. As with the results reported for conferencing and warnings, the result was adjusted to reflect the greater representation of Indigenous youths at the court level (a factor of 1.19). For conservatism, the reoffending rate for expanded diversion programs was modelled to be 32% under the intervention rather than the 25.1% estimated for conference programs. The higher rate was based on the results of the drug diversion program (Office of Crime Statistics and Research, 2012). It is recognised that the drug diversion program may not appropriately represent real world outcomes for diverted youths in the NT; however it was selected for two reasons The likelihood of success for diversion programs would depend on individual characteristics of offending youths, including the type of offence committed and their historical offending pattern. There was insufficient data to assess the impacts in the NT context, although it was considered likely that youths who proceed to court may have offending patterns that limit the effectiveness of diversion programs (e.g. they are more likely to be repeat offenders than all youths who are apprehended by police). 2. Youths may have environmental or criminogenic risk factors, or other personal factors that would limit the effectiveness of diversion programs, and would require more intensive targeted programs and additional whole of government funding to address. Some examples include factors such as family circumstances, a history of substance abuse, and intractable behaviours and/or characteristics. 21 Sensitivity analysis was conducted on reoffending rates due to the noted uncertainty around reoffending rates. There was minimal variation in the overall net benefits when changing the reoffending rates. 40

42 Benefits of changed and reduced need for supervision The benefits of diversion and bail accommodation were modelled by considering the effects of diversion on reoffending rates, and also the change in setting for the current supervision period. That is, benefits of diversion and bail accommodation were modelled for two periods. First, preventing a youth from going to court will prevent their initial supervision period in the current year, thus avoiding the entirety of the expected supervision period. A youth may then reoffend in the current year (although data are generally only available for 12 month outcomes post-intervention). There is a one-to-one relationship between the number of diversion recipients and the periods of supervision avoided in the current year based on the existing apprehension numbers. Second, reoffending benefits were then considered in the subsequent year as a change in apprehensions based on the reoffending rates for each individual diversion program. The avoided apprehensions in the subsequent year were assumed to have the same average pathways as all apprehensions in that year (i.e. the youth may end up in detention or community-based supervision, either sentenced or unsentenced based on the pathways experienced for all youths). 22 To estimate the effect of diversion programs on future reoffending, an average effectiveness was applied from Table 3.4. While it is expected that new diversion programs will need to be offered, the effectiveness of new programs will likely have the same average effectiveness at the margin when delivered to appropriate participants. Alternatively, at the margin, increased diversion eligibility may still appropriately be handled by existing programs, and the average would be applicable. For those youths who now receive diversion programs, reoffending rates are expected to be 32% within 12 months, rather than 46.5% had they gone through the court system (section ). As such, for every apprehension diverted, there would be 0.24 fewer youths apprehended in the following year. Between 2017 and 2027, 1,093 extra participants would go through NT diversion programs, resulting in 260 fewer apprehensions over the same period. The ratios of supervised youth to total apprehensions were applied to the lower number of total apprehensions to determine the subsequent changes in supervision in following years. As noted in section 2.4, of total apprehensions: 13.3% were remanded in custody; 2.7% were supervised in unsentenced community-based settings; 3.5% were sentenced to detention; 12.5% were supervised in sentenced community-based settings; and 19.0% were supervised in any setting (i.e. there is considerable overlap between the supervision groups in any given year). Thus, in the model, a change in apprehensions predicted the subsequent decrease in supervised populations. All current diversion participants were assumed to be removed from unsentenced detention and community, and sentenced community supervision orders. The benefits of diversion are shown in Table Note apprehensions are averted as the reoffending rates for youths who go through court are higher than for youths who receive diversion programs. 41

43 Table 3.5: Change in diversion participants and associated reduction in apprehensions and days of supervision in the NT Supervision type Intervention participants Additional number of participants Additional participants as % of apprehensions Apprehensions Average daily population Sentenced detention Sentenced community Unsentenced detention Unsentenced community Days of supervision ( 000s) Sentenced detention Sentenced community Unsentenced detention Unsentenced community Total Cost of supervision ($m) Sentenced detention Sentenced community Unsentenced detention Unsentenced community Total Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding Benefits of reduced crime costs and adult incarceration Diversion programs (ignoring bail accommodation for now) result in lower offending for participants, and can therefore reduce later adult incarceration due to changing the offending trajectory. Based on data from the NT Government, there were approximately 70 adults aged who were first time adult detainees and who had a juvenile detention record. The impact of additional diversion on adult detainees was modelled using average pathways and recidivism rates for youth in sentenced detention progressing to adult detention (see section 3.3 for a full description). The pathways were modelled as averages, so that less than a full sentence could be avoided in each year, and thus there can be incremental benefits over time. 42

44 Each adult sentence was assumed to have associated benefits of $345,256 in future pathways (no real growth applied) 23 as per section 2.3. For every apprehension avoided, there were expected to be $36,391 of benefits in reduced crime costs including reduced costs of crime to victims and the community as per section 2.3. As with adult incarceration costs, no real growth was applied to the cost of crimes avoided to be conservative and capture the uncertain nature of the type of crimes avoided. Table 3.6 presents the change in adult incarcerations due to increased diversionary efforts today, and the associated reduction in costs of incarceration. Table 3.6: Change in adult incarcerations and costs of crime and incarcerations due to increased diversion expenditure, real dollars, undiscounted Benefit type Reduction in adult incarcerations (n) Reduction in apprehensions (n) Reduced costs of incarcerations ($m) Reduced costs of crime ($m) Total benefits ($m) Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. The benefits of increased diversion on future crime costs and adult incarceration were expected to be around $0.44 million in 2017, increasing to around $1.17 million annually by Summary of diversion costs and benefits The benefits of increased diversion arise from additional participants being directed to appropriate programs such as conferencing, cautions, drug and alcohol programs and other new programs. The primary financial benefits are on reduced supervision costs per person in the current year i.e. it is much cheaper to provide a diversion program than it is to remand a youth in custody, or to sentence them to detention. Not only that, there are benefits of reduced recidivism modelled as lower apprehensions in the subsequent year. While these benefits are logical, they are more likely to apply at the margin, as it would be highly improbable that 100% of youths who are apprehended could be diverted each year. There are obvious costs to society of such an approach through a potential increase in crime over time and lower community safety. Thus diversion programs need to be tailored to individual characteristics and represent the varying needs of diverted youth. For example, the NT Government (2011) recognised the need for driver education programs for youth who had committed driving offences. Overall, it was estimated that benefits between 2017 and 2027 would increase from approximately $4.2 million to $6.9 million (in real discounted dollars), respectively. Over the same period, costs are expected to increase from $1.0 million to $1.5 million (in real discounted dollars). The total net benefits of diversion were estimated to be $52.2 million. 23 Data from the 2017 Report on Government Services reveal that the cost of detention per adult detainee has been stable in real terms over the last five years. 43

45 Table 3.7: Overall costs and benefits of diversion intervention, real discounted dollars (rate = 7%) Cost or benefit Benefits of diversion programs Costs of diversion programs Net benefit Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. Benefits may be higher than indicated if there are subsequent improvements in both education of young people who are successfully diverted (e.g. through education programs), and if there are then increases in lifetime earnings. From a government point of view, there would also be benefits to government through increased tax receipts and lower expenditure on the youth justice system. From a societal point of view, there would be reduced deadweight losses (economic efficiency losses imposed by raising taxes to pay for government services) associated with corrective services, and there would be benefits to individuals in the community (lower victim costs of crime). However, there would also be other costs to implement reform. Expanded diversion programs are likely to require a whole of government approach to fund initiatives (e.g. in child protection) to fully realise the benefits of greater diversion and reduce the cycle of offending. The costs that may be associated with other government expenditure have not been considered in the report. 3.2 Bail accommodation Similarly to the diversion intervention, the purpose of bail accommodation would be to reduce the number of children remanded in custody while awaiting a court finalisation outcome. The purpose would be to ensure that children attend court, and therefore it could also have potential benefits of reducing the number of children sentenced to detention for breach of justice orders. The purpose of this section is to outline the savings to the NT community for implementing bail accommodation as a suggested reform. Most young people who end up in remand rather than on bail appear not to be sent there because they are charged with serious crimes. Based on evidence from the UK, Freeman (2008) found that around a third of remandees had been sent there because of not having a fixed abode or someone willing and able to pay surety. More than a third had originally been released on bail, but had breached bail conditions. Less than a third were there because of serious offending behaviour. These rates may have some similarities in the NT context, and were used as a basis due to the lack of available evidence of similar programs in the NT. Accordingly, a third of those currently on remand would be expected to remain there (but in a secure bail accommodation facility), because of the severity of their alleged crime. The remaining remandees (excluding those who are now diverted) were assumed to receive bail accommodation support termed bail accommodation in the following tables. For the third who continue to be remanded in custody, the costs of supervision were modelled as bail accommodation as they would be remanded in a secure bail accommodation model under the proposed intervention. This group is still referred to as unsentenced detention in the following tables Costs of bail accommodation The expected costs of bail accommodation were based on information provided by the NT Government (Kerr, 2017). Territory Families has been scoping cost options to include bail accommodation in the NT. The proposed budget for Bail Support Accommodation is $15.3 million for construction costs, which would provide 58 beds in total across Darwin, Alice Springs, Tennant Creek and Katherine. Ongoing annual costs are expected to total $16.87 million per year, including: repairs and maintenance, $500,000; property management, $710,000; and 44

46 operational grants to non-government organisations, $15.66 million. As such, the ongoing annual cost of bail accommodation was expected to be $290,862 per bed, or $797 per child in bail accommodation per day. The annual operating cost was grown at the same expected real rate as for detention (i.e. 6.93% per annum) and the cost series is shown in section 2.3. This is similar to the $682 dollars per day proposed for the NT restorative detention facilities (Table 3.15) and far cheaper than current remand costs (assuming that as these are also in Don Dale, they are the same as regular detention costs). Washington DC (Department of Youth Rehabilitation Services, 2012) provides secure bail accommodation through its Youth Services Center. 24 This centre is an 88-bed secure detention facility that provides youth with 24-hour supervision, care, and custody. Services include diagnostic screenings, onsite medical care, individual and group counselling, education provided by the DC Public Schools, structured recreational activities, and family visits. Costs per youth per day in 2001 were $US 522 per day ($823). Overall, remand would be replaced with bail accommodation until court finalisation occurs. The expected costs of this model are shown in Table 3.8. While the average daily remand population in the NT was expected to be 36.9 in 2017, the average daily remand population may grow to 58.3 by As such, the total capital costs of bail accommodation were assumed to be necessary (i.e. all 58 beds) to allow for fluctuations in the average daily population numbers, noting that the total average daily population is lower with all three interventions. Table 3.8: Overall costs of bail accommodation Average daily remand population, including: Bail accommodation Unsentenced detention Costs of bail accommodation, including: Bail accommodation Unsentenced detention* Capital costs Total Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. * the costs of unsentenced detention are treated as a benefit (i.e. bail accommodation is cheaper per day compared with existing remand). Costs and participant numbers are calculated in the absence of the diversion and restorative detention interventions Benefits of bail accommodation Effectiveness of bail accommodation Bail accommodation assists young people to meet their bail conditions, so that fewer young people will be sentenced to detention for breach of justice orders. Richards and Renshaw (2013) noted that there are a limited number of evaluations of bail support programs in Australia that quantitatively estimate the benefits of the programs. As such, it is difficult to determine the efficacy of the program. The following section discusses relevant literature assessing the appropriateness of bail accommodation and support programs in Australia and internationally. Unfortunately, most of the 24 This is termed a detention centre, but in US parlance detention is used to mean remanded, and committed is used to mean sentenced. 45

47 available evidence refers to adult programs, so it was necessary to assume that programs for young people in the NT would have similar efficacy. Completion rates of bail support programs are typically positive. An evaluation of the Queensland Conditional Bail Program for youths cited a 72% successful completion rate (Venables & Rutledge, 2003). Internationally, a 2005 UK evaluation of youth national bail support schemes reported a 54% program completion rate (Thomas, 2005). An evaluation of Victoria s adult and youth Bail Support Program conducted by Henderson (2008) looked at the effects of diverting unsentenced people from prison by providing early intervention and access to drug treatment, legal welfare and housing services. Diversion was provided with the intent of improving the likelihood that the person would be granted (and successfully complete) bail. However, only 54% of the 1,720 participants successfully attended court and all subsequent treatment program episodes. The study did not provide baseline figures for comparison, but Yick (2017) shows that 53% of youth granted bail in the NT do not commit breach of bail offenses. In 2006, Deloitte UK reviewed an evaluation of the Youth Justice Agency s Bail Supervision and Support scheme, which used remand fostering 25 rather than providing bail accommodation. The majority (55%) of young people in this program fully complied with their bail supervision program. However, the recidivism rate (those that reoffend at least once within a two year period) was 91%, and subsequent offending likewise remained high, with (Deloitte UK, 2006): 40% of program participants reoffending with a less serious offence, 46% reoffending at the same level, and 14% reoffending with a more serious offence. In light of this, it was considered unlikely that youths would completely cease offending as a result of receiving bail support. However, the aim of bail support and accommodation is to improve upon the number of young people attending their court dates, and reduce the likelihood of receiving a sentence. Overall, the average completion rate for bail support was 59%, across the aforementioned studies. 26 Accordingly, the introduction of bail accommodation was assumed to result in a 59% increase in successful bail completion, and therefore reduce the subsequent breach of justice orders sentenced to detention. Approximately 17.5% of youths in sentenced detention are there for breach of justice orders (Yick, 2016), so in any given year, the average daily number of youth in detention was assumed to be reduced by 10.3%. These youths were assumed to be sentenced to community supervision instead which results in a small increase in days overall, but at a much lower cost. Table 3.9: Number of bail accommodation participants Remand, including: Bail accommodation Unsentenced detention Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. * Bail accommodation was not assumed to have a positive effect on reoffending rates for more serious cases. Benefits are calculated in the absence of the diversion and restorative detention interventions. 25 Remand fostering is an alternative use of secure accommodation when a person is under prison custody. 26 Defined as not having been re-arrested or reported to the police as being in breach of their bail supervision program, or any other bail condition, and also attending all court appearances during the bail supervision period. 46

48 Benefits of changed and reduced need for supervision As with diversion, the ratios of supervised youth to total apprehensions were applied to the lower number of total apprehensions to determine the subsequent changes in supervision in following years. The benefits of bail accommodation are shown in Table Table 3.10: Change in the number of supervised youth and funding and associated reduction in days of supervision due to bail accommodation Supervision type Average daily population Sentenced detention Sentenced community Unsentenced detention Unsentenced community* Days of supervision ( 000s) Sentenced detention Sentenced community Unsentenced detention Unsentenced community* Bail accommodation Total^ Cost of supervision ($m) Sentenced detention Sentenced community Unsentenced detention Unsentenced community* Total Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. * unsentenced community has a very small reduction in costs and days. ^ The increase in days occurs because each pathway has been modelled using the average length of stay for the changed pathway and sentenced community is slightly longer on average than sentenced detention Benefits of reduced crime costs and adult incarceration Bail accommodation reduces the number of adult sentences because some youth will be sentenced to community supervision, rather than being sentenced to detention (and then having higher subsequent recidivism). Overall, approximately 107 youths would not be sentenced to detention between 2017 and 2027, and therefore around 21 will not progress to adult detention. 47

49 Each adult sentence was assumed to have associated benefits of $345,256 in future pathways (no real growth applied) 27 as per section 2.3. Table 3.11 presents the change in adult incarcerations due to providing bail accommodation today, and the associated reduction in costs of incarceration. Table 3.11: Change in adult incarcerations and costs of crime and incarcerations due to bail accommodation, real dollars, undiscounted Benefit type Reduction in adult incarcerations (n) Reduced costs of incarcerations ($m) Source: Deloitte Access Economics calculations. Note: results are for financial years. The benefits of increased bail accommodation on future adult incarceration was expected to be around $0.5 million of additional benefits in 2017, increasing to around $0.8 million annually by Summary of bail accommodation costs and benefits The primary financial benefits of bail accommodation are reduced supervision costs in the current year i.e. based on the costs of bail accommodation proposed by the NT Government (Kerr, 2017), it is much cheaper to provide bail accommodation to youth than it is to remand them in custody, or to sentence them to detention. Overall, it was estimated that benefits between 2017 and 2027 would increase from approximately $20.7 million to $31.6 million (in real discounted dollars), respectively. Over the same period, costs are expected to increase from $7.2 million to $11.1 million (in real discounted dollars), and there would be associated upfront capital costs of $15.3 million. The total net benefits of diversion were estimated to be $183.3 million (in the absence of diversion and restorative detention). Table 3.12: Overall costs and benefits of bail accommodation, real discounted dollars (rate = 7%) Cost or benefit ($m) Benefits of bail accommodation Costs of bail accommodation Net benefit or cost Source: Deloitte Access Economics calculations. Note: results are for financial years. Components may not sum to totals due to rounding. 3.3 Restorative detention A secure accommodation model will still be required to hold children who cannot be diverted, with at least a facility in Darwin and a facility in Alice Springs. However, the facilities would be new purpose built, small scale, home style facilities in community settings. This section looks at the changes in corrective service and associated costs resulting from replacing Don Dale and related facilities with restorative detention facilities which are designed to address the factors that contribute to crime in youth, rather than simply being punitive. Despite an extensive literature search, there appear to be relatively few published journal articles outlining the cost-benefit analysis of restorative detention facilities. Three articles were identified which are outlined as follows. Latessa et al (2014) assessed the incremental costs and benefits of over 10,000 youth (matched by criminogenic risk) across three pathways in Ohio: diversion programs; Community Corrections Facilities 27 Data from the 2017 Report on Government Services reveal that the cost of detention per adult detainee has been stable in real terms over the last five years. 48

50 which are small-scale, secure community residential facilities with cognitive behaviour therapy and other therapeutic interventions; and regular youth detention. Jara (2013) examined Missouri s Multi-Systemic Treatment (MST), an intensive therapeutic intervention for chronic, violent or substance abusing juvenile offenders that houses youth in secure but small facilities. The author conducted a detailed examination of fixed and variable costs under this system, and its impact on recidivism. She then conducted a prospective CBA of a scenario where California adopted this system, assuming the same proportional reductions in recidivism. Caldwell et al (2006) reported on the Mendota Juvenile Treatment Center in Wisconsin. This facility treats violent offenders in small secure facilities, with twice the staff to resident ratio of regular youth detention, most of whom have psychological qualifications. While this study was only small scale (n=101) it had the advantage that every participant was case-matched to another youth in regular youth detention Costs of restorative detention The new facilities will have substantial capital costs at the outset, and may have higher running costs per person due to higher staff ratios, enhanced staff training, higher wages and likely higher needs of young people in detention (because they are higher risk children who are now detained). The following sections outline the expected land costs, construction costs, ongoing costs and recurrent capital costs. Largely, the costs have been based on Jara (2013), who outlines both the fixed and variable costs for MST facilities. Another restorative detention facility of the same number of beds, but specifically designed for indigenous people, the North Cheyenne Youth Services Center, has also been included for reference on the fixed costs side. Operational costs for this facility are unknown. The construction costs of a regular youth detention centre in the ACT are also included for comparison Number of facilities required Given that the average MST facility houses 36 detainees, it was necessary to determine how many facilities would be required for the NT. In section 3.2, it was assumed that most of the youths currently on remand would be sent to bail accommodation. As such, it was estimated that a facility smaller than an MST equivalent would be required. The expected average daily number of future youths sentenced to detention was not expected to rise over the next 10 years due to the benefits of restorative detention on recidivism. In most recent years the average daily number of youths in detention in the NT has been close to 50 people, and in the absence of bail accommodation it would be necessary to have sufficient capacity to account for this and expected growth in the number of youth in detention on an average day. However, given the presence of bail accommodation in the modelled intervention, the average daily number of youths in sentenced detention is expected to be closer to 15. The peak daily population of youths in detention has been up to 50% higher than average populations (for example, the annual report for the Department of Correctional Services shows that the peak number of youths in detention was 73 while the average daily number was 49 during the financial year), so facilities would need to be able to cater for around 23 youths during peak times approximately 0.64 times the size of the average MST facility Land and construction costs Jara (2013) had the most comprehensive analysis of costs, and is used as the basis for costing proposed restorative detention in the NT. Regarding establishment costs, the average MST facility is designed to hold 36 youths built upon 11.2 acres (4.5 hectares) of land. To house all of California s then current youth detainee population would require 31 such facilities. A search for rural acreage near Darwin yielded an average price of $449,450 per acre. 28 A restorative detention facility would also be needed to replace the current Alice Springs facility, however, there were not enough rural sites currently for sale near Alice Springs to gain an indication of value. As a substitute, houses 28 accessed 22 June Note, rural properties sizes are usually given in both acres and hectares. 49

51 in Alice Springs appeared to be valued around three-quarters of similar houses in Darwin. 29 Assuming that a similar ratio applies to rural land values as to house prices, and giving a double-weight to Darwin (as Don Dale usually has at least twice the number of detainees as Alice Springs) the weighted average cost of land was estimated to be $409,625 per acre. Thus, for the same amount of land as the average MST site, the total cost would be approximately $4.6 million, or $2.9 million for sufficient facilities to cater for 23 youths. Jara (2013) estimated that the average cost of acreage near urban centres in California was $US 255, Converting this to current Australian dollars (allowing for exchange rates and inflation) gives an equivalent of $443,400 which triangulates well with the similar figure for the NT. Thus land costs for the average facility would be $US 2.9 million ($4.9 million) 31. The MST centre which Jara (2013) based costs on is the New Beginnings Youth Development Center in Washington DC. This has an accommodation building, an education building, a gym, and a warehouse, which total 69,750 square feet. Californian prison construction costs were estimated at $US 241 per square foot ($418), resulting in construction costs of $US 16.8 million ($29.2 million) per facility based on Jara (2013). While the reference used by Jara was for a restorative detention facility, the costs are representative of jail construction costs in general in the United States. The Chatham County Government (2013) conducted an an analysis of 26 jails constructed in North Carolina between 2000 and 2013, ranging in size from 67 beds to 1,000 beds. The average cost per square foot was identical to that used by Jara. The Northern Cheyenne Youth Services Center is a 36 bed restorative detention facility in Minnesota. It was constructed in 2015 for $US 6.2 million ($9.3 million), or $0.2 million per bed ($0.3 million) 32. Despite the low cost, it still boasts medium security, with a strengthened perimeter fence and electronic detection systems. The facility provides 36 beds in five separate living units, an alternative school including assessment capability, probation/hearing facilities, program functions, and spaces for health, dining, recreation, as well as administrative offices. There are also separate entrances for the public and staff, supplies, and a discrete detainee entrance and processing component (Figure 3.1). Total floor space is 33,000 square feet (3,066 square metres). To help detainees prepare themselves to rejoin the wider community, it offers a wide range of work and treatment programs, and they are supervised 24/7. 33 Cultural aspects of the Northern Cheyenne Tribe are expressed in the operation and design of the Center that support traditional inter-relationships residents and staff, and incorporates traditional design forms using indigenous materials and colours. 34 There are no restorative detention centres in Australia that can be used for cost comparisons. However, the Bimberi Youth Justice Centre in the ACT, while still a secure facility, was also the first human rights compliant youth detention centre in Australia. Announced construction costs were $40 million for a 40 bed facility, or $1 million per bed accessed 22 June accessed 22 June All Jara s costs are in 2010 dollars. Converting to Australian dollars using World Bank Purchasing Power Parity for 2010, and adjusting for Australian inflation since then makes Jara s figures 74% higher. 31 Unless specified otherwise, all costs are in current Australian dollars

52 Figure 3.1: Layout of the Northern Cheyenne Youth Services Center, Montana. Source: Klein McCarthy Architects (2017). From such evidence as is available in the literature, larger restorative detention centres appear to have higher construction costs per bed, due to additional facilities such as medical and dental, or swimming pools and gymnasiums. The County of Chatham (2013) conducted a study of all the jails built in North Carolina between 2000 and 2013, which did not demonstrate a significant relationship between the cost per bed and facility size (Figure 3.2). Thus, the fact that the proposed restorative justice facilities in the Northern Territory are small does not necessarily indicate that they should be more expensive to construct than existing youth detention facilities. 51